Insulin resistance and compensatory hyperinsulinemia are characteristic features of obesity and polycystic ovary syndrome, and both are associated with reduced fertility and implantation. There is little knowledge about the effect of insulin on the decidualization process and previous findings are contradictory. We investigated the effect of insulin on the regulation of forkhead box protein O1 (FOXO1), one of the most important transcription factors during decidualization. Endometrial stromal cells were isolated from six healthy, regularly menstruating women and decidualized in vitro. Gene expression levels of six putative FOXO1 target genes (including insulin-like growth factor binding protein-1 (IGFBP1) and prolactin (PRL)) were measured with Real-Time PCR following FOXO1 inhibition or insulin treatment. PI3K inhibition was used to identify the possible mechanism behind regulation. Subcellular localization of FOXO1 was analyzed with immunofluorescence. All the genes (IGFBP1, CTGF, INSR, DCN, LEFTY2), except prolactin, were evaluated as FOXO1 target genes in decidualizing stromal cells. Insulin caused a significant dose-dependent inhibition of the verified FOXO1 target genes. It was also demonstrated that insulin regulated FOXO1 target genes by transcriptional inactivation and nuclear export of FOXO1 via PI3K pathway. However, insulin did not inhibit the morphological transformation of endometrial stromal cells via transcriptional inactivation of FOXO1. This study provides new insights on the action of insulin on the endometrium via regulation of FOXO1. It is suggested that hyperinsulinemia results in dysregulation of a high number of FOXO1 controlled genes that may contribute to endometrial dysfunction and reproductive failure. Our findings may illuminate possible reasons for unexplained infertility.
Citation: Ujvari D, Jakson I, Babayeva S, Salamon D, Rethi B, Gidlöf S, et al. (2017) Dysregulation of In Vitro Decidualization of Human Endometrial Stromal Cells by Insulin via Transcriptional Inhibition of Forkhead Box Protein O1. PLoS ONE 12(1): e0171004. https://doi.org/10.1371/journal.pone.0171004
Editor: Zeng-Ming Yang, South China Agricultural University, CHINA
Received: October 1, 2016; Accepted: January 13, 2017; Published: January 30, 2017
Copyright: © 2017 Ujvari et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: Support was provided by the Swedish Research Council (Vetenskapsrådet) 20324 to ALH [www.vr.se]; and the Stockholm County Council (Stockholms läns landsting) to ALH, SG [www.sll.se]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Insulin resistance and compensatory hyperinsulinemia are common features of obesity and polycystic ovary syndrome (PCOS), which are both associated with reduced fertility and pregnancy complications, including impaired decidualization and implantation, miscarriage, gestational diabetes, preeclampsia, intra-uterine growth restriction and preterm birth [1–6].
It is well-known that insulin resistance and hyperinsulinemia could contribute to hyperandrogenism and ovarian dysfunction . The endocrine and metabolic abnormalities associated with PCOS and obesity may also affect endometrial function and receptivity . However, very little is known about the effects of insulin on endometrial function. We have previously shown that lifestyle intervention in obese women with PCOS results in lower fasting insulin levels, enhanced insulin signaling in the endometrium and improved menstrual cyclicity .
Decidualization is differentiation of endometrial stromal fibroblasts into secretory epithelioid decidual cells that is essential for implantation and normal pregnancy. The decidual process is characterized by the expression of a variety of phenotypic markers. The most well-known genes are prolactin (PRL) and insulin-like growth factor binding protein-1 (IGFBP1) .
There is little knowledge about the effect of insulin on the decidualization process and previous findings are contradictory. It has been demonstrated that insulin down-regulates IGFBP1 in decidualizing human endometrial stromal cells [11, 12], whereas PRL production is stimulated . Thus, the effect of insulin on decidualization needs to be further explored.
The evolutionary conserved transcription factor forkhead box protein O1 (FOXO1) is one of the earliest induced transcription factors in the decidualization process of human endometrial stromal cells [10, 14, 15]. It was demonstrated that among half a thousand of genes, PRL and IGFBP1 were aberrantly expressed upon FOXO1 knockdown in decidualizing human endometrial stromal cells supporting the importance of FOXO1 during differentiation of endometrial stromal cells . Furthermore, overexpression of FOXO1 was found to induce a decidual morphology in endometrial stromal cells, suggesting a cardinal role of FOXO1 in acquisition of the epitheloid phenotype during decidualization . Previous studies have suggested that FOXO1 activity can be regulated by insulin through the PI3K pathway .
In this in vitro study, we aimed to investigate FOXO1 transcriptional activity as a possible mechanism by which insulin modulates decidualization of human endometrial stromal cells.
Materials and Methods
Six regularly cycling, non-smoking healthy volunteers underwent collection of endometrial biopsy under local anesthesia using a suction curette (Pipet Curet, CooperSurgical, USA) at cycle day 5–9. All participants were between 18–35 years and had a body mass index ranging 19–28. Hormonal medication was not allowed at least three months prior to biopsy sampling. No women had any current chronic disease, endocrine disorder, or continuous medication. One woman had had an uneventful pregnancy with normal birth at term and another woman had had one miscarriage prior to the study, she later gave birth at term after IVF treatment. No cause of the infertility could be determined. The other four women had not been pregnant. Serum levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol-17β (E2), total testosterone sex hormone-binding globulin (SHBG) and fasting glucose were determined as described previously [9, 19]. FSH (5.7±0.8 IU/l), LH (5.9±1.4 IU/l), E2 (253 (115–466 pmol/l)), total testosterone (0.9±0.3 nmol/l), SHBG (69.2±26.9 nmol/l) and fasting glucose levels (4.8±0.5 mmol/l) were in the normal range for all of them (values presented are means±SD or medians and interquartile ranges, depending on the distribution).
All women gave their written informed consent and the local ethics committee approved the study at the Regional Ethical Committee in Stockholm (Dnr 2008/865-32).
Isolation of human endometrial stromal cells
Isolation of endometrial stromal cells was carried out after modification of a method previously described [20, 21]. Briefly, the endometrial tissues were transferred in Ham’s F-10 Nut Mix medium (Thermo Fischer Scientific, USA) supplemented with 0.2% penicillin-streptomycin (Sigma-Aldrich, USA) and 20% of HI-FBS (Sigma-Aldrich, USA) on ice. After washing, the tissues were minced with a scalpel to approximately 1 mm3 pieces, and filtered through a 100 μm pore size cell strainer (Falcon BD Biosciences, Belgium) to remove red blood cells. After sedimentation (140 g, 5 min, 4°C), tissues were digested with collagenase III (Worthington, USA) (1 mg/ml) supplemented with DNase I (40 μg/ml) (Sigma-Aldrich, USA) in PBS at 37°C for 2–2.5 h in a shaking thermostat to obtain single stromal cells. After sedimentation (120g, 10 min, 4°C), the cell suspensions were washed with Ham`s F-10 Nut mix. The cell suspensions were filtered through a 40μm cell strainer (Falcon BD Biosciences, Belgium), allowing single stromal cells to pass through, while glandular tubules were restrained. Stromal cells were sedimented (120g, 10 min, 4°C) and then frozen in freezing mixture (Gibco) and kept on -150°C until use. Freezing did not affect cell viability significantly (data not shown). Cell viability was monitored using Trypan blue staining. Purity of stromal cells was ensured by sequential culturing and assessed by cytokeratin and CD10 staining for epithelial and stromal cells, respectively.
Endometrial stromal cells, 105/well, were seeded in 6-well Costar plates (Sigma-Aldrich, USA) and cultured in DMEM/F12-Glutamax (Thermo Fischer Scientific, USA) supplemented with 10% fetal bovine serum and 0.2% penicillin-streptomycin until ~80% confluency. In vitro decidualization was performed using a well-established procedure as previously described [22–24]. Briefly, media was changed to DMEM/F12 without phenol red (Thermo Fischer Scientific, USA), supplemented with 2% charcoal stripped fetal bovine serum (Sigma-Aldrich, USA) and 0.2% penicillin-streptomycin. In order to investigate the kinetics of decidualization, we treated the cells with 1 μM medroxyprogesterone-17-acetate (MPA) (Sigma-Aldrich, USA) and 1 mM N6, 2`-O-dibutyryladenosine cAMP (db-cAMP) (Sigma-Aldrich, USA) for 3 and 6 days. To clarify the contribution of each of MPA and db-cAMP to induce decidualization, endometrial stromal cells were treated with 1 μM MPA, 1 mM db-cAMP or combined treatment of MPA and db-cAMP for 6 days. Furthermore, to investigate the effect of insulin on decidualization and FOXO1 target genes, cells were treated with 1 μM MPA and 0.5 mM db-cAMP in the presence or absence of 5, 50 or 500 nM insulin (Sigma-Aldrich, USA).
To test if wortmannin, a PI3K inhibitor, can block the effect of insulin, stromal cells were pre-decidualized for 3 days and then pre-treated according to previous publications [12, 25] with 500 nM wortmannin (Sigma-Aldrich, USA) for 1 hour prior to addition of 100 nM insulin for 2 days along with decidualization agents db-cAMP and MPA. To evaluate the involvement of FOXO1 in the regulation of a particular gene, stromal cells were pre-decidualized for 3 days and then treated with 100 nM AS1842856 (Merck Millipore, Germany), a FOXO1 inhibitor, for 2 more days in the presence of decidualization agents db-cAMP and MPA. The culture media was renewed every 3 days. Depending on the experiment, cells were harvested for RNA isolation after 3, 5 or 6 days.
RNA isolation, cDNA synthesis and real-time PCR
Total RNA was extracted and 1 μg RNA was subjected to cDNA synthesis as described previously . The optimal amount of cDNA for each gene was tested with a dilution series subjected to Real-Time PCR. cDNAs of all the investigated genes were then measured with Real-Time PCR (TaqMan or SybrGreen methods) along with the housekeeping gene ribosomal protein L13a (RPL13A), using the same amount of cDNA (10–40 ng) for both target and housekeeping gene, depending on the abundance of the target gene. The employed TaqMan assays (Thermo Fischer Scientific, USA) and oligonucleotides (Sigma-Aldrich, USA) are listed in S1 and S2 Tables. The reactions were run in a StepOnePlusTM Real-Time PCR System (Thermo Fischer Scientific, USA). Gene expression levels of putative FOXO1 target genes (PRL, IGFBP1, connective tissue growth factor (CTGF), insulin receptor (INSR), decorin (DCN) and left-right determination factor 2 (LEFTY2)) and a forkhead box protein M1 (FOXM1) target gene, signal transducer and activator of transcription 3 (STAT3), known to be up-regulated upon decidualization were determined using the ΔΔCT method. The PCR efficiency with all amplicons was 90–100% and all determinations were performed in triplicate.
In order to evaluate the effect of insulin on the subcellular localization of FOXO1 and if this effect is conveyed by PI3K, endometrial stromal cells were plated into culture chambers at a density of 2x104 and cultured until ~70% confluency. The cells were pre-decidualized for 3 days and then pre-treated with 500 nM wortmannin (Sigma-Aldrich, USA) for 1 hour prior to treatment with 100 nM insulin. Since it has been published earlier that PI3K signaling led to phosphorylation, nuclear exclusion and proteasomal degradation of FOXO1 in some other cell types [26–31], we chose a shorter, 3h insulin treatment, in order to be able to capture the change in subcellular localization in response to insulin, prior to an eventual proteasomal degradation of FOXO1. After the treatment cells were fixed with 3.7% paraformaldehyde in PBS for 20 min and permeabilized by treatment with 0.1% Triton X-100 in PBS for 30 min. After blocking with 3% BSA, 0.1% Tween 20 in PBS for 1 h, cells were incubated with anti-FOXO1 antibody (Cell Signaling Technology, USA) using 1:200 dilution, followed by staining with Alexa Fluor 488 anti-mouse secondary antibody. Pictures were acquired using 630x magnification with a Zeiss AxioVert 40 CFL microscope equipped with a Zeiss AxioCAM MRM digital camera and the AxioVision software.
All values are presented as medians and ranges (min-max). Statistical analysis was performed including all the six subjects by the Friedman test followed by Dunn`s multiple comparison, where each treatment group was tested against the untreated group. In the PI3K inhibitory experiments, the insulin treated groups were compared with their respective controls (decidual, decidual + solvent or decidual + PI3K inhibitor). Normality of the data was tested with the Kolgomorov-Smirnov test. P < 0.05 was considered as significant.
db-cAMP and MPA induce decidualization in human primary endometrial stromal cells
To investigate the effect of db-cAMP and MPA on human primary endometrial stromal cells, we treated the cells with MPA, db-cAMP or their combination for 6 days. We found that db–cAMP alone induced morphological transformation of the cells and up-regulated decidualization markers, whereas MPA did not. The combined use of MPA and db-cAMP also induced morphological change of the cells and enhanced up-regulation of decidualization markers compared to db-cAMP alone (S1 Fig). Furthermore, we investigated the kinetics of decidualization in response to MPA and db-cAMP and found that decidualization was apparent after 3 days, but was even more enhanced after 6 days (S2 Fig).
Several putative FOXO1 target genes are down-regulated upon the inhibition of FOXO1 in decidualizing human endometrial stromal cells
To establish whether FOXO1 regulates PRL, IGFBP1, CTGF, INSR, DCN, and LEFTY2 gene expression levels in decidualizing endometrial stromal cells, we treated cells with a FOXO1 inhibitor (AS1842856) in the presence of the decidualizing agents MPA and db-cAMP. FOXO1 inhibition in decidualizing endometrial stromal cells resulted in down-regulation of IGFBP1 (Fold change (FC) = 0.13), CTGF (FC = 0.2), INSR (FC = 0.2), DCN (FC = 0.43) and LEFTY2 (FC = 0.005), whereas PRL (FC = 1.1) was not significantly changed (Table 1). These results indicate that FOXO1 regulates the above genes, with the exception of PRL, in human decidualizing stromal cells.
Insulin inhibits FOXO1 induced target genes in decidualizing human endometrial stromal cells
To test if insulin regulates the verified FOXO1 target genes, proliferative phase derived stromal cells were decidualized in the presence or absence of increasing concentrations of insulin, and relative gene expression levels of FOXO1 and some of its target genes were measured. FOXO1, IGFBP1, INSR, DCN and LEFTY2 were significantly up-regulated upon decidualization, while CTGF was not significantly changed (S3 Fig). On the other hand, IGFBP1, CTGF, INSR and DCN were significantly down-regulated by insulin in a dose dependent fashion, while LEFTY2 was not significantly changed (Fig 1). These results indicate that increasing concentrations of insulin in decidualizing endometrial stromal cells alter the action of FOXO1 on its target genes. In contrast, to the inhibitory effect of insulin on the verified FOXO1 target genes, PRL expression was significantly stimulated by the highest dose of insulin (FC = 3.0; P<0.05). To determine if FOXO1 itself is regulated by insulin on a transcriptional level we measured FOXO1 gene expression level in response to insulin. We found that insulin down-regulated FOXO1 gene expression level in a dose dependent fashion meaning that insulin regulates FOXO1 at a transcriptional level in decidualizing endometrial stromal cells (Fig 1). To determine if insulin has similar effect on other members of forkhead box proteins of importance for decidualization of endometrial stromal cells, we measured the gene expression level of STAT3, a well-known target of FOXM1. We found that STAT3 was significantly up-regulated upon decidualization, but was not changed by any of the concentrations of insulin, suggesting that the action of insulin is specific to FOXO1 in these cells (S4 Fig).
Relative gene expression levels of IGFBP1, CTGF, INSR, DCN, LEFTY2, and FOXO1 in response to 5, 50 and 500 nM insulin in in vitro decidualized human endometrial stromal cells after 6 days. The values presented are medians and ranges (min-max). * = p < 0.05; ** = p < 0.01; *** = p < 0.001 in comparison to the control value.
Inhibition of FOXO1 is insufficient to prevent morphological transformation of endometrial stromal cells during decidualization
We also investigated the decidualization of endometrial stromal cells with microscopy and flow cytometry in the absence or presence of insulin. We found that despite the significant downregulation of several FOXO1 target genes, insulin did not inhibit the morphological transformation of spindle-shaped endometrial stromal cells into polygonal epitheloid-like cells (cell size: stromal vs decidual P < 0.05; stromal vs decidual+insulin P < 0.05; decidual vs decidual+insulin not significant (NS); cell granularity: stromal vs decidual P < 0.05; stromal vs decidual+insulin P = 0.0551 (NS); decidual vs decidual+insulin P = 0.0782 (NS)) (Fig 2).
Representative micrographs of undifferentiated stromal and decidualized cells (MPA + db-cAMP) in the absence or presence of insulin were taken using an inverted microscope with 40x magnification. Forward vs side scatter (FCS/SSC) dot plot generated by flow cytometer showing the cell distribution of undifferentiated stromal and decidualized cells in the absence or presence of insulin.
Insulin regulates FOXO1 transcriptional activity through PI3K in decidualizing human endometrial stromal cells
To investigate if insulin regulates FOXO1 through the PI3K pathway we blocked PI3K signaling with wortmannin during the decidualization process in the presence or absence of insulin. It was found that PI3K inhibition partially reversed the insulin-induced down-regulation of IGFBP1, CTGF, INSR, DCN and LEFTY2 (Fig 3). However, gene expression levels of PRL and FOXO1 were not significantly influenced by PI3K inhibition (Fig 3). These results suggest that regulation of FOXO1 transcriptional activity by insulin is conveyed by PI3K signaling. It must be noted, that PI3K inhibition in the absence of insulin did not have any significant effect on the investigated genes (not shown), meaning that PI3K inhibition did not affect the expression of these genes at this stage in decidualizing endometrial stromal cells.
Relative levels of IGFBP1, CTGF, INSR, DCN, LEFTY2 and FOXO1 in response 500 nM wortmannin, a PI3K inhibitor in combination with 100 nM insulin in in vitro decidualized human endometrial stromal cells after 4 days. The values presented are medians and ranges (min-max). * = p < 0.05; ** = p < 0.01; in comparison to the appropriate control value.
Insulin induces nuclear export of FOXO1 in decidualizing human endometrial stromal cells
We investigated the subcellular localization of FOXO1 in decidualizing endometrial stromal cells in response to insulin treatment by fluorescent microscopy. In the absence of insulin, FOXO1 was predominantly localized within the nucleus (Fig 4). In contrast, insulin treatment rapidly induced the nuclear export of FOXO1 (Fig 4). To assess if insulin regulation of nuclear export of FOXO1 is signaled through PI3K we blocked this pathway in decidualizing endometrial stromal cells in the presence or absence of insulin. We found that PI3K inhibition blocked insulin’s effect on nuclear FOXO1 export, suggesting that insulin induces nuclear export of FOXO1 through PI3K signaling (Fig 4).
Representative fluorescent micrographs for FOXO1 (green) and DAPI-stained nucleus (blue) in decidualized cells with 630x magnification. FOXO1 is located predominantly in the nucleus in the absence of insulin, while exported from the nucleus in the presence of insulin. Wortmannin, a PI3K inhibitor has no effect on the subcellular localization of FOXO1 alone, while in combination with insulin it blocks insulin induced nuclear export of FOXO1.
This study is the first to show that insulin alters gene expression of certain decidual markers  and other genes, which are up-regulated upon decidualization by regulation of FOXO1 transcriptional activity. It was demonstrated by the determination of mRNA expression that insulin regulates FOXO1 target genes through transcriptional inactivation and nuclear export of FOXO1 in decidualizing primary human endometrial stromal cells. Furthermore, insulin regulates FOXO1 activity not only on a post-translational level, but also to a minor extent on a transcriptional level. However, insulin does not inhibit the morphological transformation of endometrial stromal cells to decidualized cells.
We demonstrated that MPA and db-cAMP induced decidualization of endometrial stromal cells in agreement with previous observations [22–24]. Furthermore, we evaluated the ability and contribution of MPA and db-cAMP to induce decidualization and found that MPA alone did not induce decidualization markers, but enhanced the effect of db-cAMP remarkably as shown previously [10, 32, 33].
We confirmed IGFBP1 and LEFTY2 as FOXO1 target genes , and further demonstrated the regulation of CTGF, INSR and DCN by FOXO1, but not PRL, in decidualizing endometrial stromal cells. Previous findings on PRL regulation by FOXO1 in decidualizing human endometrial stromal cells have been inconclusive. Buzzio et al showed an up-regulation of PRL by FOXO1 silencing upon decidualization . In contrast, Takano et al reported that silencing of FOXO1 with siRNA resulted in significant down-regulation of PRL in decidualizing endometrial stromal cells . Thus, there is no clear evidence for a consistent regulation of PRL by FOXO1 in these cells.
Here we further show that insulin dose-dependently down-regulates the FOXO1 target genes IGFBP1, CTGF, INSR and DCN, which are up-regulated upon decidualization. In contrast, gene expression of PRL was up-regulated by insulin. Our results showing different regulation of IGFBP1 and PRL by insulin are in agreement with others findings. It was previously demonstrated that IGFBP1 was clearly down-regulated in decidualizing endometrial stromal cells in response to insulin [11, 12], while PRL was slightly enhanced . However, investigating the decidual progression in response to insulin using solely PRL and IGFBP1 as markers fails to draw attention to the possible regulation of these genes via transcriptional inactivation of FOXO1 . This study shows that insulin may influence a high number of genes that are up-regulated upon decidualization via transcriptional inactivation of FOXO1, whereas PRL seems to be regulated by insulin through a different mechanism.
We found no significant difference in morphological transformation of spindle-shaped stromal cells to polygonal epithelial-like decidual cells by decidualization agents db-cAMP and MPA in the presence or absence of insulin, meaning that insulin does not inhibit decidualization morphologically. Hypothetically this could be explained by an action of PRL. Previous publications suggested a critical role of PRL in the differentiation of a variety of cell types including endometrial glandular epithelial cells [34–41]. Furthermore, PRL was shown to act as an inducer of decidualization of endometrial stromal cells [42, 43]. We could confirm others’ observation on insulin induced up-regulation of PRL expression in decidual cells . It is therefore suggested that increased expression of PRL, as a differentiation factor could be responsible for the morphological transformation of stromal cells in the presence of insulin. However, the exact role of prolactin in the induction of decidualization markers and morphological transformation of endometrial stromal cells has to be investigated further.
Inhibition of the PI3K pathway resulted in a partial reversion of the effect of insulin on the FOXO1 target genes, suggesting that insulin regulates FOXO1 transcriptional activity mainly by PI3K signaling. The most likely explanation for the incomplete inhibition is that insulin also can induce PI3K independent signaling cascades to regulate the investigated FOXO1 target genes. Alternatively, insulin signaling may also modulate the activity of the protein binding partners of FOXO1, which specifically can either activate or repress FOXO1 target genes [15, 18, 44–52].
We took into consideration the regulation of other forkhead box proteins by insulin in decidualizing endometrial stromal cells. Forkhead box protein O3A (FOXO3A) was shown to be down-regulated upon decidualization, whereas forkhead box protein O4 (FOXO4) is not expressed in the endometrium . On the other hand, we confirmed the up-regulation of STAT3 upon decidualization , but failed to show its regulation by insulin, suggesting that the effect of insulin was specific to FOXO1. It was also shown that inhibition of FOXM1 could prevent decidualization , while insulin could not in our experiments. These observations further suggest that insulin does not inhibit the activity of FOXM1 in decidualizing stromal cells.
External stimuli control FOXO activity primarily by regulating its subcellular localization . In normal luteal phase endometrium, cAMP induces nuclear accumulation of FOXO1 protein, while continuous progesterone signaling induces nuclear export of FOXO1. The latter is resulting in transcriptional inactivation of a fraction of the total FOXO1 pool. Fall of progesterone levels in the late luteal phase of the menstrual cycle rapidly leads to re-accumulation of nuclear FOXO1, expression of pro-apoptotic FOXO1 target genes, cell death and menstrual bleeding [33, 56]. It has been demonstrated in a number of other cell types, that when Akt, downstream of PI3K, is activated the FOXO proteins are directly phosphorylated [26–31]. This results in their transcriptional inactivation through their nuclear export, followed by a multistep negative regulation including ubiquitination and proteasomal degradation . Previous studies have also suggested that FOXO1 activity can be regulated by insulin through the PI3K pathway (19) (Fig 5).
A model on FOXO1 regulation through insulin-induced phosphorylation-dependent degradation as suggested in some other cell types: PI3K is activated by insulin, the FOXO proteins are phosphorylated, resulting in their transcriptional inactivation through their nuclear export and cytoplasmic retention followed by a multistep negative regulation including ubiquitination and proteasomal degradation [26–31].
Our study is the first to demonstrate that insulin induces transcriptional inactivation of FOXO1 by its nuclear export via PI3K pathway in decidualizing endometrial stromal cells, as shown in other cell types . Although post-translational inhibition of FOXO proteins has been well defined in a large diversity of cells, other levels of regulation, such as mRNA expression, has remained largely unexplored. In this study, we were able to show that insulin also down-regulates FOXO1 mRNA expression in decidualizing endometrial stromal cells. Very limited data is available on transcriptional down-regulation of FOXO1 by insulin, however it was previously demonstrated that both transcriptional and post-translational regulation share the PI3K pathway [57, 58].
Finely tuned decidualization is critical to normal embryo implantation, placentation and maintenance of pregnancy. Disturbances in this process and dysregulation of genes induced upon decidualization may lead to both reduced fertility and adverse pregnancy outcomes, as it was already shown for most of those factors we investigated here [59–63]. Our study suggests an adverse effect of insulin on FOXO1 regulation during decidualization. Furthermore, despite ovulatory cycles or ovulation induction, hyperinsulinemia in women with PCOS and obesity may still affect the endometrium adversely. Insulin may negatively affect the nuclear re-accumulation of FOXO1 prior to menstruation that might lead to dysregulation of FOXO1 induced pro-apoptotic genes. This may contribute to endometrial dysfunction and menstrual disturbances in women with hyperinsulinemic conditions such as PCOS and obesity [64–66]. We have previously shown that reduced insulin levels in obese women with PCOS after lifestyle intervention resulted in improved endometrial insulin signaling and menstrual cyclicity . This indirectly supports the role of FOXO1 regulation in endometrial breakdown. Our present findings highlight the importance of an increased knowledge of the action of insulin on endometrial function, and may illuminate possible reasons to unexplained infertility. It has to be noted that our aim was to study basic molecular mechanisms associated with hyperinsulinemia only, and not in the context of specific disorders. However, we do expect the general mechanisms to be applicable to many hyperinsulinemic disorders.
In summary, this study is the first to show that insulin can down-regulate the expression of FOXO1 induced genes in decidualizing primary human endometrial stromal cells. Insulin regulates FOXO1 activity predominantly on a post-translational level via its nuclear export and to a minor extent on a transcriptional level. We have shown that insulin action on FOXO1 activity in decidualizing endometrial stromal cells is mainly mediated through the PI3K pathway. This study suggests that insulin dysregulates FOXO1 induced genes upon decidualization, although morphological transformation is not inhibited.
S1 Fig. Relative gene expression levels in response to MPA, db-cAMP and their combined treatment.
A. Representative micrographs of undifferentiated stromal cell and cells in response to MPA (1 μM), db-cAMP (0.5 mM) and their combined treatment after 6 days were taken using an inverted microscope with 40x magnification. B. Relative gene expression levels of IGFBP1, CTGF, INSR, DCN, LEFTY2 and FOXO1 in response to MPA (1 μM), db-cAMP (0.5 mM) and their combined treatment in endometrial stromal/decidualizing cells after 6 days. The values presented are medians and ranges (min-max). * = p < 0.05 and ** = p < 0.01 in comparison to the control (stromal) value.
S2 Fig. Relative gene expression levels in response to combined treatment of MPA and db-cAMP after 3 and 6 days.
A. Representative micrographs of undifferentiated stromal and decidualized cells (MPA + db-cAMP) after 3 and 6 days were taken using an inverted microscope with 40x magnification. B. Relative gene expression levels of IGFBP1, CTGF, INSR, DCN, LEFTY2 and FOXO1 in response to decidualization agents MPA (1 μM) and db-cAMP (0.5 mM) in endometrial stromal/decidual cells after 0, 3 and 6 days. The values presented are medians and ranges (min-max). * = p < 0.05, ** = p < 0.01 and *** = p < 0.001 in comparison to the control (stromal) value.
S3 Fig. Relative gene expression levels in response to combined treatment of MPA and db-cAMP after 6 days.
Relative gene expression levels of IGFBP1, CTGF, INSR, DCN, LEFTY2 and FOXO1 in the absence or presence of decidualization agents MPA (1 μM) and db-cAMP (0.5 mM) in endometrial stromal/decidual cells after 6 days. The values presented are medians and ranges (min-max). * = p < 0.05 and *** = p < 0.001 in comparison to the control value.
S4 Fig. Relative gene expression level of STAT3 in response to combined treatment of MPA and db-cAMP and insulin.
A. Relative gene expression level of STAT3 in the absence or presence of decidualization agents MPA (1 μM) and db-cAMP (0.5 mM) in endometrial stromal/decidual cells after 6 days. The values presented are medians and ranges (min-max). * = p < 0.05 in comparison to the control value. B. Relative gene expression level of STAT3 in response to decidualization agents MPA (1 μM) and db-cAMP (0.5 mM) in the presence or absence of 5, 50 and 500 nM insulin in endometrial stromal cells after 6 days. The values presented are medians and ranges (min-max).
S1 Table. TaqMan assay applied in Real-Time PCR.
TaqMan assays applied for amplification of prolactin (PRL), insulin-like growth factor binding protein-1 (IGFBP1) and insulin receptor (INSR). RPL13A was used as an endogenous control.
S2 Table. Oligonucleotides applied in Real-Time PCR.
Forward and reverse oligos applied for amplification of connective tissue growth factor (CTGF), decorin (DCN) and left-right determination factor 2 (LEFTY2). RPL13A was used as an endogenous control.
The authors thank Lotta Blomberg, Siv Rödin Andersson and Berit Legerstam at the Women`s Health Clinical Research Unit for technical assistance.
- Conceptualization: DU IJ DS SG ALH.
- Formal analysis: DU IJ DS BR.
- Funding acquisition: ALH.
- Investigation: DU IJ SB DS BR.
- Methodology: DU IJ DS BR.
- Resources: ALH.
- Supervision: DU SG ALH.
- Validation: DU IJ DS BR.
- Visualization: DU IJ DS BR.
- Writing – original draft: DU IJ.
- Writing – review & editing: DU IJ SB DS BR SG ALH.
- 1. Jung SJ, Park SK, Shin A, Lee SA, Choi JY, Hong YC, et al. Body mass index at age 18–20 and later risk of spontaneous abortion in the Health Examinees Study (HEXA). BMC pregnancy and childbirth. 2015;15:228. Pubmed Central PMCID: 4582827. pmid:26403183
- 2. Katulski K, Czyzyk A, Podfigurna-Stopa A, Genazzani AR, Meczekalski B. Pregnancy complications in polycystic ovary syndrome patients. Gynecological endocrinology: the official journal of the International Society of Gynecological Endocrinology. 2015 Feb;31(2):87–91.
- 3. O'Brien TE, Ray JG, Chan WS. Maternal body mass index and the risk of preeclampsia: a systematic overview. Epidemiology. 2003 May;14(3):368–74. pmid:12859040
- 4. Roos N, Kieler H, Sahlin L, Ekman-Ordeberg G, Falconer H, Stephansson O. Risk of adverse pregnancy outcomes in women with polycystic ovary syndrome: population based cohort study. Bmj. 2011;343:d6309. Pubmed Central PMCID: 3192872. pmid:21998337
- 5. Solomon CG, Seely EW. Brief review: hypertension in pregnancy: a manifestation of the insulin resistance syndrome? Hypertension. 2001 Feb;37(2):232–9. pmid:11230277
- 6. Wolf M, Sandler L, Munoz K, Hsu K, Ecker JL, Thadhani R. First trimester insulin resistance and subsequent preeclampsia: a prospective study. The Journal of clinical endocrinology and metabolism. 2002 Apr;87(4):1563–8. pmid:11932283
- 7. Diamanti-Kandarakis E, Dunaif A. Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocrine reviews. 2012 Dec;33(6):981–1030. pmid:23065822
- 8. Giudice LC. Endometrium in PCOS: Implantation and predisposition to endocrine CA. Best practice & research Clinical endocrinology & metabolism. 2006 Jun;20(2):235–44.
- 9. Ujvari D, Hulchiy M, Calaby A, Nybacka A, Bystrom Ba, Hirschberg AL. Lifestyle intervention up-regulates gene and protein levels of molecules involved in insulin signaling in the endometrium of overweight/obese women with polycystic ovary syndrome. Human reproduction. 2014 Jul;29(7):1526–35. pmid:24842895
- 10. Gellersen B, Brosens J. Cyclic AMP and progesterone receptor cross-talk in human endometrium: a decidualizing affair. The Journal of endocrinology. 2003 Sep;178(3):357–72. pmid:12967329
- 11. Giudice LC, Dsupin BA, Irwin JC. Steroid and peptide regulation of insulin-like growth factor-binding proteins secreted by human endometrial stromal cells is dependent on stromal differentiation. The Journal of clinical endocrinology and metabolism. 1992 Nov;75(5):1235–41. pmid:1385468
- 12. Lathi RB, Hess AP, Tulac S, Nayak NR, Conti M, Giudice LC. Dose-dependent insulin regulation of insulin-like growth factor binding protein-1 in human endometrial stromal cells is mediated by distinct signaling pathways. The Journal of clinical endocrinology and metabolism. 2005 Mar;90(3):1599–606. pmid:15613433
- 13. Thrailkill KM, Golander A, Underwood LE, Richards RG, Handwerger S. Insulin stimulates the synthesis and release of prolactin from human decidual cells. Endocrinology. 1989 Jun;124(6):3010–4. pmid:2656248
- 14. Brar AK, Handwerger S, Kessler CA, Aronow BJ. Gene induction and categorical reprogramming during in vitro human endometrial fibroblast decidualization. Physiological genomics. 2001 Dec 21;7(2):135–48. pmid:11773600
- 15. Christian M, Zhang X, Schneider-Merck T, Unterman TG, Gellersen B, White JO, et al. Cyclic AMP-induced forkhead transcription factor, FKHR, cooperates with CCAAT/enhancer-binding protein beta in differentiating human endometrial stromal cells. The Journal of biological chemistry. 2002 Jun 7;277(23):20825–32. pmid:11893744
- 16. Takano M, Lu Z, Goto T, Fusi L, Higham J, Francis J, et al. Transcriptional cross talk between the forkhead transcription factor forkhead box O1A and the progesterone receptor coordinates cell cycle regulation and differentiation in human endometrial stromal cells. Molecular endocrinology. 2007 Oct;21(10):2334–49. pmid:17609436
- 17. Buzzio OL, Lu Z, Miller CD, Unterman TG, Kim JJ. FOXO1A differentially regulates genes of decidualization. Endocrinology. 2006 Aug;147(8):3870–6. pmid:16690806
- 18. Barthel A, Schmoll D, Unterman TG. FoxO proteins in insulin action and metabolism. Trends in endocrinology and metabolism: TEM. 2005 May-Jun;16(4):183–9. pmid:15860415
- 19. Nybacka A, Carlstrom K, Stahle A, Nyren S, Hellstrom PM, Hirschberg AL. Randomized comparison of the influence of dietary management and/or physical exercise on ovarian function and metabolic parameters in overweight women with polycystic ovary syndrome. Fertility and sterility. 2011 Dec;96(6):1508–13. pmid:21962963
- 20. Arnold JT, Kaufman DG, Seppala M, Lessey BA. Endometrial stromal cells regulate epithelial cell growth in vitro: a new co-culture model. Human reproduction. 2001 May;16(5):836–45. pmid:11331626
- 21. Satyaswaroop PG, Bressler RS, de la Pena MM, Gurpide E. Isolation and culture of human endometrial glands. The Journal of clinical endocrinology and metabolism. 1979 Apr;48(4):639–41. pmid:429509
- 22. Frolova AI, O'Neill K, Moley KH. Dehydroepiandrosterone inhibits glucose flux through the pentose phosphate pathway in human and mouse endometrial stromal cells, preventing decidualization and implantation. Molecular endocrinology. 2011 Aug;25(8):1444–55. Pubmed Central PMCID: 3146244. pmid:21680659
- 23. Matsui N, Kawano Y, Nakamura S, Miyakawa I. Changes in vascular endothelial growth factor production associated with decidualization by human endometrial stromal cells in vitro. Acta obstetricia et gynecologica Scandinavica. 2004 Feb;83(2):138–43. pmid:14756729
- 24. Yuhki M, Kajitani T, Mizuno T, Aoki Y, Maruyama T. Establishment of an immortalized human endometrial stromal cell line with functional responses to ovarian stimuli. Reproductive biology and endocrinology: RB&E. 2011;9:104. Pubmed Central PMCID: 3160358.
- 25. Zeng GY, Quon MJ. Insulin-stimulated production of nitric oxide is inhibited by wortmannin—Direct measurement in vascular endothelial cells. J Clin Invest. 1996 Aug 15;98(4):894–8. English. pmid:8770859
- 26. Biggs WH 3rd, Meisenhelder J, Hunter T, Cavenee WK, Arden KC. Protein kinase B/Akt-mediated phosphorylation promotes nuclear exclusion of the winged helix transcription factor FKHR1. Proceedings of the National Academy of Sciences of the United States of America. 1999 Jun 22;96(13):7421–6. Pubmed Central PMCID: 22101. pmid:10377430
- 27. Brownawell AM, Kops GJ, Macara IG, Burgering BM. Inhibition of nuclear import by protein kinase B (Akt) regulates the subcellular distribution and activity of the forkhead transcription factor AFX. Molecular and cellular biology. 2001 May;21(10):3534–46. Pubmed Central PMCID: 100275. pmid:11313479
- 28. Kops GJ, de Ruiter ND, De Vries-Smits AM, Powell DR, Bos JL, Burgering BM. Direct control of the Forkhead transcription factor AFX by protein kinase B. Nature. 1999 Apr 15;398(6728):630–4. pmid:10217147
- 29. Matsuzaki H, Daitoku H, Hatta M, Tanaka K, Fukamizu A. Insulin-induced phosphorylation of FKHR (Foxo1) targets to proteasomal degradation. Proceedings of the National Academy of Sciences of the United States of America. 2003 Sep 30;100(20):11285–90. Pubmed Central PMCID: 208749. pmid:13679577
- 30. Rena G, Guo S, Cichy SC, Unterman TG, Cohen P. Phosphorylation of the transcription factor forkhead family member FKHR by protein kinase B. The Journal of biological chemistry. 1999 Jun 11;274(24):17179–83. pmid:10358075
- 31. Takaishi H, Konishi H, Matsuzaki H, Ono Y, Shirai Y, Saito N, et al. Regulation of nuclear translocation of forkhead transcription factor AFX by protein kinase B. Proceedings of the National Academy of Sciences of the United States of America. 1999 Oct 12;96(21):11836–41. Pubmed Central PMCID: 18373. pmid:10518537
- 32. Fan W, Li SW, Li WH, Wang Y, Gong Y, Ma QH, et al. FOXO1 expression and regulation in endometrial tissue during the menstrual cycle and in early pregnancy decidua. Gynecologic and obstetric investigation. 2012;74(1):56–63. pmid:22538365
- 33. Labied S, Kajihara T, Madureira PA, Fusi L, Jones MC, Higham JM, et al. Progestins regulate the expression and activity of the forkhead transcription factor FOXO1 in differentiating human endometrium. Molecular endocrinology. 2006 Jan;20(1):35–44. pmid:16123151
- 34. Horseman ND, Zhao WZ, Montecino-Rodriguez E, Tanaka M, Nakashima K, Engle SJ, et al. Defective mammopoiesis, but normal hematopoiesis, in mice with a targeted disruption of the prolactin gene. Embo J. 1997 Dec 1;16(23):6926–35. English. pmid:9384572
- 35. Jabbour HN, Critchley HOD, Boddy SC. Expression of functional prolactin receptors in nonpregnant human endometrium: Janus kinase-2, signal transducer and activator of transcription-1 (STAT1), and STAT5 proteins are phosphorylated after stimulation with prolactin. J Clin Endocr Metab. 1998 Jul;83(7):2545–53. English. pmid:9661641
- 36. Jabbour HN, Critchley HOD, Yu-Lee LY, Boddy SC. Localization of interferon regulatory factor-1 (IRF-1) in nonpregnant human endometrium: Expression of IRF-1 is up-regulated by prolactin during the secretory phase of the menstrual cycle. J Clin Endocr Metab. 1999 Nov;84(11):4260–5. English. pmid:10566682
- 37. Krasnow JS, Hickey GJ, Richards JS. Regulation of Aromatase Messenger-Rna and Estradiol Biosynthesis in Rat Ovarian Granulosa and Luteal Cells by Prolactin. Molecular endocrinology. 1990 Jan;4(1):13–21. English. pmid:1970119
- 38. Lahat N, Miller A, Shtiller R, Touby E. Differential-Effects of Prolactin Upon Activation and Differentiation of Human B-Lymphocytes. J Neuroimmunol. 1993 Aug;47(1):35–40. English. pmid:8376547
- 39. McAveney KM, Gimble JM, YuLee LY. Prolactin receptor expression during adipocyte differentiation of bone marrow stroma. Endocrinology. 1996 Dec;137(12):5723–6. English. pmid:8940406
- 40. Murphy WJ, Durum SK, Longo DL. Differential-Effects of Growth-Hormone and Prolactin on Murine T-Cell Development and Function. J Exp Med. 1993 Jul 1;178(1):231–6. English. pmid:8315380
- 41. Nevalainen MT, Valve EM, Ahonen T, Yagi A, Paranko J, Harkonen PL. Androgen-dependent expression of prolactin in rat prostate epithelium in vivo and in organ culture. Faseb J. 1997 Dec;11(14):1297–307. English. pmid:9409549
- 42. Tabanelli S, Tang BQ, Gurpide E. Invitro Decidualization of Human Endometrial Stromal Cells. J Steroid Biochem. 1992 May;42(3–4):337–44. English.
- 43. Wu WX, Brooks J, Glasier AF, Mcneilly AS. The Relationship between Decidualization and Prolactin Messenger-Rna and Production at Different Stages of Human-Pregnancy. J Mol Endocrinol. 1995 Apr;14(2):255–61. English. pmid:7619213
- 44. Dowell P, Otto TC, Adi S, Lane MD. Convergence of peroxisome proliferator-activated receptor gamma and Foxo1 signaling pathways. The Journal of biological chemistry. 2003 Nov 14;278(46):45485–91. pmid:12966085
- 45. Hirota K, Daitoku H, Matsuzaki H, Araya N, Yamagata K, Asada S, et al. Hepatocyte nuclear factor-4 is a novel downstream target of insulin via FKHR as a signal-regulated transcriptional inhibitor. The Journal of biological chemistry. 2003 Apr 11;278(15):13056–60. pmid:12519792
- 46. Kim JJ, Buzzio OL, Li S, Lu Z. Role of FOXO1A in the regulation of insulin-like growth factor-binding protein-1 in human endometrial cells: interaction with progesterone receptor. Biology of reproduction. 2005 Oct;73(4):833–9. Pubmed Central PMCID: 1237025. pmid:15987820
- 47. Kim JJ, Taylor HS, Akbas GE, Foucher I, Trembleau A, Jaffe RC, et al. Regulation of insulin-like growth factor binding protein-1 promoter activity by FKHR and HOXA10 in primate endometrial cells. Biology of reproduction. 2003 Jan;68(1):24–30. pmid:12493691
- 48. Mahmud DL, M GA , Deb DK, Platanias LC, Uddin S, Wickrema A. Phosphorylation of forkhead transcription factors by erythropoietin and stem cell factor prevents acetylation and their interaction with coactivator p300 in erythroid progenitor cells. Oncogene. 2002 Feb 28;21(10):1556–62. pmid:11896584
- 49. Nasrin N, Ogg S, Cahill CM, Biggs W, Nui S, Dore J, et al. DAF-16 recruits the CREB-binding protein coactivator complex to the insulin-like growth factor binding protein 1 promoter in HepG2 cells. Proceedings of the National Academy of Sciences of the United States of America. 2000 Sep 12;97(19):10412–7. Pubmed Central PMCID: 27038. pmid:10973497
- 50. Puigserver P, Rhee J, Donovan J, Walkey CJ, Yoon JC, Oriente F, et al. Insulin-regulated hepatic gluconeogenesis through FOXO1-PGC-1alpha interaction. Nature. 2003 May 29;423(6939):550–5. pmid:12754525
- 51. Schuur ER, Loktev AV, Sharma M, Sun Z, Roth RA, Weigel RJ. Ligand-dependent interaction of estrogen receptor-alpha with members of the forkhead transcription factor family. The Journal of biological chemistry. 2001 Sep 7;276(36):33554–60. pmid:11435445
- 52. Zhao HH, Herrera RE, Coronado-Heinsohn E, Yang MC, Ludes-Meyers JH, Seybold-Tilson KJ, et al. Forkhead homologue in rhabdomyosarcoma functions as a bifunctional nuclear receptor-interacting protein with both coactivator and corepressor functions. The Journal of biological chemistry. 2001 Jul 27;276(30):27907–12. pmid:11353774
- 53. Kajihara T, Jones M, Fusi L, Takano M, Feroze-Zaidi F, Pirianov G, et al. Differential expression of FOXO1 and FOXO3a confers resistance to oxidative cell death upon endometrial decidualization. Molecular endocrinology. 2006 Oct;20(10):2444–55. pmid:16709600
- 54. Jiang Y, Liao Y, He H, Xin Q, Tu Z, Kong S, et al. FoxM1 Directs STAT3 Expression Essential for Human Endometrial Stromal Decidualization. Scientific reports. 2015 Sep 03;5:13735. Pubmed Central PMCID: 4558550. pmid:26334131
- 55. Calnan DR, Brunet A. The FoxO code. Oncogene. 2008 Apr 7;27(16):2276–88. pmid:18391970
- 56. Kajihara T, Brosens JJ, Ishihara O. The role of FOXO1 in the decidual transformation of the endometrium and early pregnancy. Medical molecular morphology. 2013 Jun;46(2):61–8. pmid:23381604
- 57. Essaghir A, Dif N, Marbehant CY, Coffer PJ, Demoulin JB. The transcription of FOXO genes is stimulated by FOXO3 and repressed by growth factors. The Journal of biological chemistry. 2009 Apr 17;284(16):10334–42. Pubmed Central PMCID: 2667720. pmid:19244250
- 58. Hinman RM, Bushanam JN, Nichols WA, Satterthwaite AB. B cell receptor signaling down-regulates forkhead box transcription factor class O 1 mRNA expression via phosphatidylinositol 3-kinase and Bruton's tyrosine kinase. Journal of immunology. 2007 Jan 15;178(2):740–7.
- 59. Irwin JC, Giudice LC. Insulin-like growth factor binding protein-1 binds to placental cytotrophoblast alpha5beta1 integrin and inhibits cytotrophoblast invasion into decidualized endometrial stromal cultures. Growth hormone & IGF research: official journal of the Growth Hormone Research Society and the International IGF Research Society. 1998 Feb;8(1):21–31.
- 60. Salker M, Teklenburg G, Molokhia M, Lavery S, Trew G, Aojanepong T, et al. Natural selection of human embryos: impaired decidualization of endometrium disables embryo-maternal interactions and causes recurrent pregnancy loss. PloS one. 2010;5(4):e10287. Pubmed Central PMCID: 2858209. pmid:20422017
- 61. Siddiqui MF, Nandi P, Girish GV, Nygard K, Eastabrook G, de Vrijer B, et al. Decorin over-expression by decidual cells in preeclampsia: a potential blood biomarker. American journal of obstetrics and gynecology. 2016 Sep;215(3):361 e1– e15.
- 62. Tang M, Naidu D, Hearing P, Handwerger S, Tabibzadeh S. LEFTY, a member of the transforming growth factor-beta superfamily, inhibits uterine stromal cell differentiation: a novel autocrine role. Endocrinology. 2010 Mar;151(3):1320–30. Pubmed Central PMCID: 2840686. pmid:20056823
- 63. Waddell JM, Evans J, Jabbour HN, Denison FC. CTGF expression is up-regulated by PROK1 in early pregnancy and influences HTR-8/Svneo cell adhesion and network formation. Human reproduction. 2011 Jan;26(1):67–75. Pubmed Central PMCID: 3005999. pmid:21098624
- 64. Bouzas IC, Cader SA, Leao L, Kuschnir MC, Braga C. Menstrual cycle alterations during adolescence: early expression of metabolic syndrome and polycystic ovary syndrome. Journal of pediatric and adolescent gynecology. 2014 Dec;27(6):335–41. pmid:25256874
- 65. Falsetti L, Eleftheriou G. Hyperinsulinemia in the polycystic ovary syndrome: a clinical, endocrine and echographic study in 240 patients. Gynecological endocrinology: the official journal of the International Society of Gynecological Endocrinology. 1996 Oct;10(5):319–26.
- 66. Weiss DJ, Charles MA, Dunaif A, Prior DE, Lillioja S, Knowler WC, et al. Hyperinsulinemia is associated with menstrual irregularity and altered serum androgens in Pima Indian women. Metabolism: clinical and experimental. 1994 Jul;43(7):803–7.